Academic journal article Memory & Cognition

Not the Same Old Hindsight Bias: Outcome Information Distorts a Broad Range of Retrospective Judgments

Academic journal article Memory & Cognition

Not the Same Old Hindsight Bias: Outcome Information Distorts a Broad Range of Retrospective Judgments

Article excerpt

The hindsight bias (e.g., Fischhoff, 1975) illustrates that outcome information can make people believe that they would have (or did) predict an outcome that they would not (or did not) actually predict. In two experiments, participants (N = 226) made a prediction immediately before receiving outcome information. Therefore, participants could not distort or misremember their predictions to make them align with the outcome information. In both experiments, participants distorted their reports of how certain they recalled having been in their prediction, how good of a basis they had for making the prediction, how long they took to make the prediction, and so forth. Experiment 2 showed that these effects were diminished when participants engaged in private thought about the upcoming questions prior to receiving outcome information, suggesting that the effect is not due to impression management concerns.

A robust psychological phenomenon known as the hindsight bias or "knew it all along" effect provides a framework for determining how people respond to outcome information (Fischhoff, 1975; Fischhoff & Beyth, 1975; Wood, 1978).' In typical demonstrations of the hindsight bias, people learn outcome information and then indicate what they would have predicted had they not been given the outcome information (e.g., Fischhoff's [1977] hypothetical paradigm). The hindsight bias exists when people overestimate their ability to have predicted the outcome. The hindsight bias also exists when people misremember earlier predictions so that they conform with outcome information (e.g., Fischhoff's [1977] memory paradigm). The typical hindsight paradigms do not account for situations in which people make predictions, receive outcome information, and accurately recall their prior predictions. The present research is designed to address just such a situation. In doing so, the experiments presented below address the effect of outcome information on reports of the decision processes that accompany a prediction rather than on reports of the prediction itself.

Research on the Hindsight Bias

The hindsight bias appears in two related experimental paradigms. For example, Fischhoff( 1977) had people answer general knowledge questions (e.g., "Absinthe is [a] a precious stone or [b] a liqueur") in one of three groups. In the memory group, people answered questions, received the correct answers, and later recalled their responses. In the reliability group, people answered questions and then later recalled their responses (without knowledge of the correct answers). In the hypothetical group, people received information about the correct answers and then indicated how they would have answered the questions, had they not known the correct answers. People in the hypothetical group overestimated how much they would have known, compared with the reliability group's original responses, a phenomenon Fischhoff termed the "I knew it all along" effect. Fischhoff (1977) also observed a bias in the memory group. People who misremembered their earlier judgments did so in such a way that their recalled judgments were closer to the correct answer than were their original judgments. Therefore, there are two situations in which people display the hindsight bias. In the first, outcome information distorts people's estimates of how they would have answered, had they not known the outcome information (i.e., Fischhoff's hypothetical group). In the second, outcome information distorts people's recall of prior judgments (i.e., Fischhoff's memory group). Early research demonstrated that the hindsight bias was not a function of misinterpreted instructions (i.e., as a display of current knowledge instead of retrospective judgments; Fischhoff & Beyth, 1975) or demand characteristics (Wood, 1978).

Since Fischhoff's initial experiments, the hindsight bias has been demonstrated many times in both memory and hypothetical experimental paradigms (see metaanalysis by Christensen-Szalanski & Willham, 1991; for a review, see Hawkins & Hastie, 1990). …

Search by... Author
Show... All Results Primary Sources Peer-reviewed

Oops!

An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.